Skip to main content

Full text of "Guide to the analysis of potable spirits"

See other formats


IS OF 



POTABLE SPIRITS 



S. A, VASEY 



IN 

Edmund 




GUIDE TO THE ANALYSIS OF 
POTABLE SPIRITS 



GUIDE TO THE ANALYSIS 



OF 



POTABLE SPIRITS 



T *. 






BY 



S. ARCHIBALD VASEY, F.I.C., F.C.S. 

MEMBER OF THE SOCIETY OF PUBLIC ANALYSTS 




LONDON 

BAILLlfeRE, TINDALL AND COX 

8, HENRIETTA STREET, COVENT GARDEN 

1904 

[All rights reserved] 



IN HEMOR1AM 

' ' 



PREFACE 

THE object of the present work is to induce analysts 
to take up a more detailed examination of potable 
spirits than has hitherto been their practice, and 
to bring analytical methods to bear more minutely 
upon this question, in order that further light may 
be thrown on the extent of substitution which is 
going on in connection with the sale of spirits. 
The Sale of Food and Drugs Act ought to protect 
the purchaser of spirits in the same way as it pro- 
tects the purchaser of foods, who is entitled to be 
supplied with an article of the nature, substance, 
and quality demanded ; and for the purposes of the 
Act the expression " food " includes every article 
used for food or drink by man other than drugs or 
water. The author believes that the methods de- 
scribed in the following pages will, with a little 
practice, enable the analyst to declare whether a 
given spirit submitted to him is or is not genuine 
according to its description. Certain of tbe pro- 
cesses recounted are based upon those employed in 



vi PREFACE 

the Municipal Laboratory of Paris, and the author 
is indebted to Messrs. Masson et Cie., of Paris, for 
permission to use some of the descriptive matter 
and statistics contained in MM. Girard and Cuni- 
asse's " Manuel Pratique de F Analyse des Alcohols 
et des Spiritueux," published by them. To his col- 
league and friend, Dr. H. P. Cholmeley, the author 
owes considerable assistance in the translation of 
some of the technical sections of this work. The 
methods which have been introduced by analysts 
in this country, notably Allen, Hewitt, and Schid- 
rowitz, have been embodied. The names of the 
authors and the references to original papers of the 
processes referred to are acknowledged in the text. 
The author has recorded his own experiences of the 
processes as applied to the discrimination of spirits, 
and has given what he has found to be in practice 
the procedure which in his hands has yielded con- 
cordant and trustworthy results. During a visit 
to Cognac recently he had the advantage of making 
the acquaintance of M. Ordonneau, whose classical 
inquiries on alcohol and spirits are well known, and 
the author takes this opportunity of expressing his 
thanks to M. Ordonneau for the valuable informa- 
tion which he gained while in his laboratory. As 
far as the author is aware, there is no treatise in 



PREFACE vii 

English that is exclusively devoted to the subject, 
and in most manuals on the analysis of foods and 
drugs, including malt liquors, wines, and spirits, the 
question of differentiating one kind of spirit from 
another is either ignored entirely or receives such 
scant treatment as to afford no practical guidance 
whatever. In endeavouring to fill this breach, the 
author would urge upon public analysts to give 
greater practical study than heretofore to this in- 
teresting and important branch of analytical work. 
The question is bound to occupy their attention 
sooner or later in connection with their official 
duties. Already several convictions have been ob- 
tained throughout the country under the Sale of 
Food and Drugs Act, in which the evidence of 
analysis was accepted as proving certain samples of 
" brandy " to be not genuine grape spirit. 

S. ARCHIBALD VASEY. 

BEECHCROFT, 

BURNT ASH LANE, 
BROMLEY, KENT. 



CONTENTS 

CHAPTER PACK 

I. INTRODUCTORY - 1 1 

II. THE CHEMICAL CHARACTERISTICS OF POTABLE 

SPIRITS - -15 

III. METHODS OF ANALYSIS I STANDARDS, AND THE 

USE OF PURE ALCOHOL IN ANALYSIS ; 
DISTILLATION OF THE SPIRIT FOR ANALYSIS 21 

IV. THE ESTIMATION OF ALDEHYDES AND ACIDS - 28 
V. THE ESTIMATION OF THE HIGHER ALCOHOLS - 36 

VI. THE ESTIMATION OF COMPOUND ETHERS - 57 

VII. THE ESTIMATION OF FURFURAL - 60 

VIII. THE IMPORTANCE OF TASTE - 67 

IX. GENERAL CONSIDERATIONS OF THE RESULTS OF 

ANALYSIS - 71 

X. COLORIMETERS - "79 

APPENDIX I TABLES GIVING TYPES OF GRAIN 
AND ROOT SPIRITS, GENUINE WHISKIES, 
BRANDIES, RUM AND GIN, BLENDED SPIRITS 82 



IX 



CHAPTER I 

INTRODUCTORY 

THE analysis of spirits has been much neglected in 
this country, partly because, Jittle attempt has been 
made to draw up a systematic scheme for their 
examination, but -more ,. pa jticuiajly, because no 
further control over the sale of ardent spirits has 
been exercised beyond that which relates to their 
alcoholic strength. 

It is quite obvious that under such a regime it 
matters little whether a given spirit be called 
brandy, whisky, rum, or gin ; the public analyst 
would record any single member of the series as 
genuine so long as it conformed to a certain 
alcoholic strength. This is not satisfactory, nor 
is it in the public interest, when we consider that, 
properly speaking, brandy should be the produce 
of distilled wine, whisky of malt, and rum of sugar, 
and so on, each spirit possessing its own special 
characteristics. 

ii 



12 THE ANALYSIS OF POTABLE SPIRITS 

It should not be difficult, therefore, one would 
think, to define each spirit with regard to its origin ; 
and the impression of the public is that, when 
brandy is asked for, grape spirit is supplied, while 
whisky is generally understood to be a spirit ob- 
tained from barley malt, and rum a spirit distilled 
from cane-sugar or molasses. The fact that the 
strictly genuine article is rarely supplied, except it 
be very specially requested, is due to the existence 
of much cheaper source? #f akohol than are grapes, 
malt, or cane-sugar,' and to much more economical 
methods'- ol ^Vadiicihg: rt : Alcohol can be obtained 
from almost anything nowadays, even refuse of all 
kinds, and the patent still would seem to be capable 
of purifying it sufficiently to make it drinkable 
and suitable for blending purposes. The introduc- 
tion of grain or neutral spirit was never demanded 
by the public taste : it was dictated solely and 
simply by economical motives. The time required 
for maturing genuine whisky is thus saved, and the 
practice of blending is carried to an enormous, if not 
appalling, extent, with handsome profits to the 
blenders. There is no control over the character or 
purity of the grain spirit used for this purpose ; it 
may be made from any loathsome material, so long as 
it is capable of alcoholic fermentation; Clearly the 



INTRODUCTORY 13 

public interest and health need guarding here, and 
there is every reason why that clause of the Food 
and Drugs Act providing that an article should be 
of the nature, quality, and substance demanded by 
the purchaser should be enforced in the case of the 
sale of spirits. It is a serious thing that when 
brandy i.e., genuine grape spirit is demanded, a 
much inferior spirit is supplied pure grain spirit, 
perhaps, flavoured with French plums, " oil of 
Cognac," or even a little rum possessing none of 
the medicinal qualities of genuine brandy. 

Attempts have already been made to amend the 
law relating to the sale of spirits, and recently a 
Bill was presented by Sir Herbert Maxwell to Par- 
liament, providing for the marking of casks and 
other vessels containing whisky. The memorandum 
of the Bill is as follows : 

" The object of this Bill is to secure to purchasers 
of whisky a clear statement whether it is a whisky 
made from barley malt alone, or is in part a spirit 
made from unmalted grain. To carry out this 
object, the Bill proposes to enact that all whisky 
shall, from the time of leaving the distillery till sold 
to the consumer, be described by a mark or label 
on the cask or bottle as * malt whisky,' or ' grain 
whisky,' as the case may be. Provision is made for 



14 THE ANALYSIS OF POTABLE SPIRITS 

similar information being given to persons purchas- 
ing whisky on draught. 

" Under the existing law the Inland Revenue 
authorities are furnished with returns of the 
materials used in each distillation, and they have, 
by means of the permits or certificates required on 
the removal of spirits, and the stock-books required 
to be kept by rectifiers, dealers, and retailers, the 
information necessary for tracing the whisky till 
it reaches the retailer. Power to prosecute for 
offences against the Bill is therefore given to officers 
of Inland Revenue, but persons authorized by local 
authorities having the execution of the laws relating to 
food and drugs are also to have power to prosecute, 
and to be supplied by the Inland Revenue with the 
information necessary for that purpose." (The 
italics are the author's.) 



CHAPTER II 

THE CHEMICAL CHARACTERISTICS OF 
POTABLE SPIRITS 

THE patent or fractionating still is practically the 
key to the situation as regards the analysis of 
potable spirits. As is well known to chemists 
and distillers, this ingenious apparatus produces 
a very strong and practically featureless, because 
pure, spirit. It separates certain by-products of 
fermentation during the distillation process, with 
the result that when the spirit is highly rectified 
it is very pure and free from characteristic odour 
and flavour. The simplest form of still, or the 
pot-still, which is used in the production of brandy, 
whisky, and rum, yields, on the other hand, a more 
or less impure spirit, owing partly to the formation 
of secondary products of fermentation, and partly 
because of bodies formed during distillation, which, 
coming over with the spirit, impart to it a flavour 
and odour characteristic of the material used in the 
fermentation. 

15 



16 THE ANALYSIS OF POTABLE SPIRITS 

It is only in the pot-still spirit that these products 
occur to any extent ; they are nearly, if not en- 
tirely, eliminated in patent spirit, according to the 
degree of rectification. It is thus necessary that 
pot-still spirits should be matured by age, while 
with patent spirit no such maturing process is 
called for, unless, owing to partial rectification, 
there is a small percentage of by-products present. 
The use of patent spirit, therefore, obviates to a 
large extent the expense of storage, which is neces- 
sary for mellowing the raw spirit of the pot-still 
to render it, not only drinkable, but palatable 
and of agreeable flavour. Flavour is essential to 
potable spirits, and patent spirit, on account of its 
tastelessness, must be flavoured, which is generally 
accomplished by adding a small proportion of old, 
well-matured pot-still spirit to it. Foreign flavour- 
ings can also doubtless be used, and there is much 
brandy on the market which is absolutely destitute 
of grape spirit, just as there is much whisky that is 
absolutely destitute of malt spirit. 

It would be superfluous, of course, to use barley 
malt or grape juice for the fermentation when the 
separation of the spirit is consigned to the patent 
still. To begin with, these materials are much more 
expensive sources of alcohol than are potato starch, 



CHEMICAL CHARACTERISTICS 17 

maize, beetroot ; and, secondly, the flavours charac- 
teristic of the grape or barley malt would be to a 
great extent lost in the patent still, and, in fact, the 
product would resemble patent spirit produced from 
grain or root. Plain rectified spirit is no doubt 
made in France and elsewhere from unsound wines, 
or wines that would not make good brandy, the 
patent still being used to eliminate undesirable 
by-products. 

It is evident that pot-still spirit cannot be dis- 
pensed with so long as flavour is looked for of the 
kind which has characterized brandy and whisky 
from the earliest time they were made. Pot-still 
spirit, whether brandy or whisky, is the agent which 
the blender adds to featureless spirit in order to 
give it a characteristic flavour the genuine flavour 
of grape or malt spirit, as the case may be. To 
some extent, just as grain or patent spirit is poor 
as regards flavouring matters, so is it nearly, if not 
entirely, free from substances other than alcohol 
from the chemical point- of view it is, in fact, a 
much purer spirit than pot-still spirit. On the 
other hand, pot-still spirit is more or less strongly 
flavoured according to the material used, and chemi- 
cal analysis would seem to indicate that flavour is 
coincident with the presence of secondary products 

2 



18 THE ANALYSIS OF POTABLE SPIRITS 

arising from interaction, fermentation, or distilla- 
tion, amongst which may be reckoned compound 
ethers, volatile acids, aldehydes (including furfural), 
and alcohols of the higher series e.g., amylic and 
butylic alcohol. 

There is little doubt that the secondary or by- 
products of a pot-still spirit comprise other products 
not present in grain or patent spirit, but those just 
enumerated are, in the author's experience, suffi- 
cient to determine by chemical analysis whether a 
spirit conforms to the description given of it. The 
premises are, then, that a grain or patent spirit is 
free from, or contains a relatively small proportion 
of, by-products, or impuretes, as the French chemists 
call them, while a pot-still spirit as, for example, 
brandy made from grapes, whisky from barley malt, 
rum from sugar or molasses contains these im- 
puretes in relatively large amounts, and in a vary- 
ing ratio to each other, according to the origin of 
the spirit. It is thus not difficult to distinguish by 
means of chemical analysis pot-still spirit from 
patent spirit. 

When a potable spirit is found to be chemically 
featureless, it indicates, as a rule, an all-grain spirit, 
or the presence of a large proportion of grain or 
patent spirit, and its flavour is feeble or " delicate " 



CHEMICAL CHARACTERISTICS 19 

when compared with a pot-still spirit or an all-malt 
or all-grape spirit. That being so, the results of a 
chemical analysis based upon an accurate deter- 
mination of the secondary or by-products, or im- 
puretes, gives a very fair indication of the nature of 
the spirit. To illustrate this (see also Tables VII., 
VIII., IX., X., XL), some typical analyses made by 
the author are given in the following table. The 
results are given in grammes per 100 litres of abso- 
lute alcohol present. 



2 2 



20 



THE ANALYSIS OF POTABLE SPIRITS 



* 8~ 


_ M _ 


co O 


tx 


lift! 


ON tx O 


cs ON 
co -^ 


00 

VO 


|i 


o o o 

p * 90 


P $ 


O 
9 




00 00 


ON O 

o\ a> 


00 
(N 







CO 


"> 


ntl* 


\S) VO ^O 


? .8 


8 


g jj ^ O 


tX >H 


O N 


o 


OOQ^ 




1-1 M 


CO 


ssliisi^ 


O O 
oo O r3 


co 


CO 



2 4s ^-^ ^ "^ -S. 


vb b d 


00 


vo 


^^^Is^ 011 " 






CO 


Jshtsi 


P 8 8 


oo O 
vo O 


00 

oo 


^ * o S *^P x 


OO Li- ^ 


ON O 


W"i 


^\ "o r^ *\ ti] ^ 


"* M 


oo O 


VO 


oa< s 




M 


CO 


|.| 


oo 


oo O 


8 




a M a 


tx -t 


CO 


3 




CO Tj- 


oo 


~~"S ^^ 


O O "^ 


O <u 




IH^Il 


7*" P^ r 

oo * O 


oo o 

CO ^ 


;<* 
tx 

CO 












r 


"rt 






M 


rt 


en 




K^T? 


*-i en 


3 


s 


11 


t'l 


1 


"^* 


% <n 


o 




05 


CO *J3 O 

3! 11 

13 2 ! 

> < fo 


OT ^ 

Gj 

1 1 

<^5 .SP 

5 w 


a 
s 



CHAPTER III 
METHODS OF ANALYSIS 

I. STANDARDS, AND THE USE OF PURE ALCOHOL 
IN ANALYSIS 

PURE alcohol is a sine qua non in the preparation of 
all standard solutions, and for the purpose of making 
up solutions to a definite volume. That is to say, 
the alcohol used in the colorimetric processes, as 
well as in the preparation of standard solutions, 
must be free from aldehyde, furfural, and higher 
alcohols. In the author's experience, rectified 
spirits of wine or methylated spirit " denatured " 
simply by wood spirit is better than absolute 
alcohol, and its use is certainly more economical. 
This will be found to be comparatively free from 
aldehydes, furfural, and higher alcohols, and to 
give negative reactions (which should be ascer- 
tained) with the colour reagents, while the ordinary 
absolute alcohol frequently contains aldehydes. 
Ethers, if present, may be removed by distillation 

21 



22 THE ANALYSIS OF POTABLE SPIRITS 

after saponification with potash, and aldehydes, 
including furfural, by means of Hewitt's reagent 
(Journal of the Society of Chemical Industry, Janu- 
ary 31, 1902), the sodium salt of phenyl-hydrazine 
parasulphonic acid. MM. Girard and Cuniasse add 
to every litre of alcohol 3 or 4 grammes of chloro- 
hydrate of metaphenylenediamine, or the same 
quantity of phosphate of aniline, leaving the mix- 
ture for some days, or just boiling it for an hour 
under a reflux condenser, and finally distilling off 
the alcohol slowly, rejecting the first few c.c., and 
stopping the process as soon as nine-tenths of the 
alcohol has passed to the receiver. It will be 
found that Hewitt's reagent, employed in the same 
way, answers admirably, and the addition of meta- 
phenylenediamine is satisfactory when a distillate 
free from furfural is desired. The author, however, 
has found no difficulty in procuring rectified spirits 
of wine, which needs no preliminary purifying 
treatment on either of the lines indicated. An 
abundant stock of dilute alcohol (50 per cent, 
strength) should be available when colorimetric 
determinations are to be made. 

In ordinary practice the author has found the 
following strengths of solutions for colorimetric 
analysis convenient 5 Aldehyde, 0*05 gramme per 



METHODS OF ANALYSIS 23 

litre ; furfural, 0*01 gramme per litre ; and iso- 
butylic alcohol, 2 grammes per litre. Stronger 
solutions in 50 per cent, spirit should be kept in 
stock, from which to make the dilute solutions. 
They should be kept in the dark in non-actinic 
glass bottles. A solution of aldehyde may be 
standardized by starting with ammonia aldehyde, 
as described in the chapter on Aldehydes. There 
is no difficulty in obtaining pure furfural or iso- 
butylic alcohol (or amyl alcohol) from dealers in 
fine chemicals Messrs. Baird and Tatlock, for 
example, or Merck, of Darmstadt. The pure sub- 
stances should be weighed in flasks containing 
alcohol. 

In order to make the conditions of experiment as 
far as possible the same, the best plan is, after 
making up the standards with 50 per cent, alcohol, 
to put the solutions in the still, and treat them in 
the same way as the spirit under examination i.e., 
to distil off nine-tenths, and make up the distillate 
with water. Thus, .200 c.c. of standard furfural 
made up with pure spirit of 50 per cent, strength 
are distilled until there is 1 80 c.c. of distillate. The 
distillation is then stopped, and the distillate made 
up to 200 c.c. with water. The alcoholic strength 
will be 50 per cent. The standard so obtained 



24 THE ANALYSIS OF POTABLE SPIRITS 

should be marked " distilled standard," and its 
strength as regards furfural may be taken to be 
the same as the standard before distilling. By 
this plan the spirit would be assayed under pre- 
cisely the same conditions of experiment. Similar 
steps may be taken with the standard aldehyde and 
higher alcohol, though they may be considered to 
be superfluous. At any rate, such precautions 
would be excessive in the case of colourless spirits, 
which are at the same time free from extractive 
matters. Such, however, with, perhaps, the excep- 
tion of unsweetened gin, hardly exist amongst 
potable spirits. 



II. DISTILLATION OF THE SPIRIT FOR 
ANALYSIS 

For colorimetric analysis the spirit must be dis- 
tilled, in order to separate the volatile by-products 
from colouring matters and extractives. It would 
appear to be an advantage if in these colorimetric 
determinations the distillation of the spirit could 
be dispensed with. Hewitt and Schidrowitz have 
tried removing the colour with lead acetate, etc. 
(see p. 34), so that the spirit after nitration from 
the precipitated colouring matters can be submitted 



METHODS OF ANALYSIS 25 

directly to the colorimetric process. By adopting 
standards, however, distilled under the same con- 
ditions as the potable spirit under examination, any 
error arising from distillation is for all practical 
purposes eliminated. P. Schidrowitz (Journal of 
the Society of Chemical Industry, June 30, 1902) 
recommends a method of steam distillation for the 
separation of colouring and extractive matters. 
The distilling flask is provided with a cork covered 
with tinfoil, through which passes, to within a few 
millimetres of the bottom of the flask, a glass tube 
suitably bent and connected with a steam supply. 
First, roughly, nine-tenths of the liquid is dis- 
tilled by means of a gas-burner in the ordinary way, 
and then steam is turned on, and the burner under 
the flask so regulated that the contents of the same 
are reduced to about 5 c.c. by the time 150 c.c. in 
all have passed over. The author does not find 
any advantage in this method over simple distilla- 
tion. It is not necessary to carry the distillation 
much further than when alcohol ceases to come 
over, for it will be found that the ethers, higher 
alcohols, aldehydes, and furfural in the original 
spirit come over with the first runnings, which are 
of high alcoholic strength. This laboratory experi- 
ence is in accord with that of the distillery. Further, 



26 THE ANALYSIS OF POTABLE SPIRITS 

if the distillation be pushed too far, there is risk of 
furfural being formed in the experiment, and of 
other decompositions taking place. 

Dr. Hewitt (Journal of the Society of Chemical 
Industry, January 31, 1902) states that, since dis- 
tillation cannot be carried to dryness, on account 
of charring the residue, and furfural being of a high 
boiling-point the boiling-point of furfural is 161 C. 
all the furfural will not be obtained in the dis- 
tillate by one distillation. He proposes, therefore, 
to distil rapidly " to nearly the last drop," then to 
add 10 c.c. of pure methylated spirit and distil 
again, and then another 10 c.c., and yet a third 
10 c.c. In this way the residue in the flask will be 
found to yield no reaction for furfural. The author 
finds, however, that a solution containing 0*01 fur- 
fural per 100 c.c. of 50 per cent, alcohol yields the 
whole of the furfural in 40 c.c. of the distillate. 
The boiling-point of furfural, when it occurs in 
such small quantities as in spirits, does not affect 
its being carried over just as is the case with essential 
oils ; a parallel case also is that of the higher 
alcohols, which will be found in the first fraction 
of the distillate, though their boiling-point is con- 
siderably higher, of course, than ethylic alcohol. 
The higher alcohols, as well as the ethers, will be 



METHODS OF ANALYSIS 27 

found to be in considerable excess in the first 
runnings of the pot-still in the distillation of both 
brandy and whisky. Furfural, it may be noted, is 
very readily formed on the mere distillation of a 
weak solution of saccharine matter. Thus, Schiff 
found that by distilling 0*00005 gramme of sugar, a 
distillate is obtained giving a marked furfural reaction . 

In the writer's experience, rapid distillation until 
nine-tenths of the spirit has passed into the receiver 
is all that is required, and the results are accurate. 
The residue in the flask from the distillation of 
200 c.c. of spirit may be reserved for the examina- 
tion for extractives, sugar and tannin. 

For the estimation of aldehydes, furfural, com- 
pound ethers, and higher alcohols, 200 c.c. of the 
spirit should be distilled until 180 c.c. are collected 
in the distillate. It should then be made up to 
200 c.c. with pure distilled water at 15*5 C., and 
thoroughly mixed, and the specific gravity deter- 
mined either by weighing a known volume or by 
means of the Westphal balance or hydrometer. 
The alcoholic strength may thus be incidentally 
determined. The distillate is then reserved for 
the determination of aldehydes, furfural, ethers, 
and higher alcohols, according to the methods 
described in the following chapters. 



CHAPTER IV 
ESTIMATION OF ALDEHYDES AND ACIDS 

I. ALDEHYDES 

ALDEHYDES exist probably for the most part in 
brandy, whisky, and rum as acetaldehyde CH 3 COH, 
but there is also furfuraldehyde in small quantity 
present. The aldehydes in general have their origin in 
the oxidation of corresponding alcohols, while furfural 
is a product of the decomposition of carbohydrates. 
Pentoses are formed from woody fibre, and these, 
when heated with dilute mineral acid solution, give 
furfural. The proportion of aldehydes in potable 
spirit, however, of the CnH 2 nO series is infinitely 
greater than that of furfural, and in the estimation 
of aldehydes by the colorimetric process about to 
be described the furfural may be disregarded. The 
reagent introduced by Guyon and Schiff (Comptes 
Rendus de V Academic des Sciences, 1887, p. 1182) for 
the estimation of aldehydes is prepared in the follow- 
ing manner : A solution of fuchsine is prepared by 

28 



ESTIMATION OF ALDEHYDES AND ACIDS 29 

dissolving i gramme in a litre of water, and also a 
solution of bisulphite of soda of a strength corre- 
sponding to a specific gravity of 1360 ; 150 c.c. of the 
fuchsine solution and 100 c.c. of the bisulphite solu- 
tion are added to a litre of pure distilled water, and 
finally 15 c.c. of pure sulphuric acid (66 per cent.) are 
added. The mixture is shaken, and after some hours it 
should be quite clear and colourless. Unless a great 
number of analyses are being conducted, it is advis- 
able to keep the solutions separate, and to mix the 
quantity of each to make sufficient reagent just when 
required. Generally, it will be found convenient to 
take 5 or 10 c.c. of the spirit (50 per cent, alcohol) 
for the estimation, and to add 4 c.c. of bisulphite 
fuchsine reagent, and make up to 20 c.c. with 50 per 
cent, alcohol. In most cases (brandy and whisky) 
5 c.c. of the spirit will suffice, made up to 20 c.c. 
with 50 per cent, alcohol. The standard-aldehyde is 
made up in one or other of the following ways. 
MM. Girard and Cuniasse direct that the aldehyde 
should be freshly prepared, for it must be remem- 
bered that this body polymerizes with great facility, 
and paraldehyde has not the same action upon the 
bisulphite' reagent. Pure acetic aldehyde boils at 
20-8 C., and has a density of 0791 at 15 C. The 
standard solution may be made up by weight, or as 



30 THE ANALYSIS OF POTABLE SPIRITS 

containing so many grammes of pure aldehyde per 
volume of alcohol of 50 per cent, strength. In the 
former case the following is an example given by MM. 
Girard and Cuniasse. If 1*922 grammes of aldehyde 
are taken to make a solution of i gramme per 100 c.c., 
the volume of the solution should be 192*2 c.c. The 

I*Q22 

aldehyde occupies a volume of -~ = 2*40 c.c. and 

192*2 c.c. - 2*40 c.c. = 189*8 c.c. of alcohol to be added, 
or 177*3 grammes of alcohol at 50 (specific gravity 
0*934). The solution, then, should weigh 177*3 
grammes + 1*922 grammes = 179* 22 grammes. The 
contents of the flask whose weight is known must be 
brought to 179*22 grammes with alcohol of 50 percent, 
strength. The resulting solution would be a i per 
cent, by weight solution of aldehyde. In weighing out 
the aldehyde, it is best to drop it into the flask con- 
taining some alcohol from a pipette, so as to avoid loss 
in weighing. The solution of aldehyde in alcohol is 
permanent, but it is a good precaution to preserve the 
solution in a non-actinic glass bottle in a dark cup- 
board. MM. Girard and Cuniasse observe that the 
same procedure in regard to the preparation of a stan- 
dard aldehyde solution may be followed in making up 
the other standard solutions. By adding alcohol of 
50 per cent, strength to them, weak standard solu- 



ESTIMATION OF ALDEHYDES AND ACIDS 31 

tions may be prepared according to requirement. 
The advantage of making weight for weight standard 
solutions is doubtful ; it is preferable to have a given 
weight of substance dissolved in a known volume of 
alcohol at 50 per cent, strength, and it is convenient 
to record the results of analysis in grammes per 
hectolitre of alcohol, or, what is the same thing, in 
milligrammes per 100 c.c. 

There must always be doubt as to the percentage 
of real aldehyde contained in the specimens supplied 
by the makers. The strength of a standard solution 
of aldehyde, however, may be insured by following 
the procedure adopted by M. Rocques, who takes 
pure ammonia aldehyde as the starting-point of the 
standard aldehyde solution. This substance may 
very readily be prepared in the pure state in the 
laboratory, but the commercial product answers the 
purpose quite well. Advantage is taken of the fact 
that aldehyde ammonia is soluble in alcohol, and is 
decomposed into a corresponding salt and aldehyde 
by dilute acid. The ammonia aldehyde is first 
ground in a mortar with ether, and the ether de- 
canted several times. After drying in a current of 
air and then in a vacuum over concentrated sul- 
phuric acid, 1*386 grammes of the ammonia aldehyde 
are weighed out. This quantity corresponds to 



32 THE ANALYSIS OF POTABLE SPIRITS 

I gramme of aldehyde. The salt is perfectly definite, 
having the formula CH 3 CHOHNH 3 . The salt is 
placed in a small flask graduated to 100 c.c., and dis- 
solved in 50 c.c. of alcohol of 96 per cent, strength. 
To this are added 227 c.c. of normal alcoholic sul- 
phuric acid, which throws out ammonium sulphate. 
The volume is made up to the 100 c.c. mark with 
alcohol of 96 per cent, strength and an addition of 
0*8 c.c. of alcohol, to compensate for the volume of 
the ammonium sulphate. The mixture is shaken 
and allowed to stand for twenty-four hours, and then 
filtered. It represents an alcoholic solution of 
aldehyde containing i per cent., which can be diluted 
to the strength convenient for the estimation (viz., 
0*05 or o'io gramme of aldehyde per litre of alcohol 
at 50 per cent, strength). Solutions cf aldehyde 
made directly from aldehyde may be standardized by 
means of this solution. 

In general, 5 c.c. to 10 c.c. of the distillate of a 
potable spirit of 50 per cent, alcoholic strength is 
sufficient for the estimation of aldehyde. To this 
quantity in a graduated cylinder is added 4 c.c. of 
bisulphite of fuchsine solution, and the whole made 
up to the 20 c.c. mark. Five c.c. of the standard 
solution are treated in the same way, and the two 
cylinders are left in the cold for twenty minutes, 



ESTIMATION OF ALDEHYDES AND ACIDS 33 

when the tints are compared. The maximum in- 
tensity is developed in this time. By matching the 
two solutions, the amount of aldehyde in the spirit 
under examination may be determined. This can 
be done by decanting one or other of the coloured 
fluids until the columns are the same tint and noting 
the relative volumes. (See chapter on Colori- 
meters.) 

P. Schidrowitz states that the estimation of the 
aldehydes in whisky by the above process has one 
serious drawback (Journal of the Society of Chemical 
Industry, June 30, 1902), which is that when the dis- 
tilled liquid is treated with the reagent an opal- 
escence or turbidity is produced which renders the 
colorimetric estimation difficult. The author has 
never encountered this drawback, which may pos- 
sibly arise from employing too large a volume of 
distillate (5 c.c. is sufficient in general), or a distillate 
of a higher alcoholic strength than 50 per cent. 
However, Schidrowitz was led to modify the process 
after observing that when -the reagent was added to 
an undistilled sample of a colourless whisky the 
liquid remained perfectly bright. He ultimately 
obtained satisfactory results by partially decolour- 
izing the spirit (complete decolourization appears to 
be impossible by means of precipitants), and colouring 

3 



34 THE ANALYSIS OF POTABLE SPIRITS 

slightly the standard solution in imitation of the 
sample. The substance chosen for partially decolour- 
izing the spirit was basic lead acetate, removing the 
excess by means of a saturated solution of potassium 
sulphate. Neither of these reagents influence the 
colour produced by the rosaniline reagent, as was 
shown by experiment. The control solution is 
coloured with either tincture of galls or the colouring 
matter derived from evaporating a genuine cask- 
coloured spirit to dryness on the water-bath, and 
taking up with pure 50 per cent, alcohol. Caramel 
is inadmissible for the purpose, as, unlike the other 
colourings suggested, it influences the colour reac- 
tion. This process has the decided advantage that 
any loss or change of aldehydes due to distillation 
is avoided. 

II. VOLATILE AND FIXED ACIDITY. 

The acidity, both volatile and fixed, is best deter- 
mined separately by titrating a known volume of the 
original spirit (25 c.c.) with decinormal solution of 
baryta, which gives the total acidity. A further 
25 c.c. is evaporated to dryness three or four times 
on the water-bath, and the residue taken up with 
water, and titrated as before with decinormal 
baryta, using in both cases phenolphthalein as 



ESTIMATION OF ALDEHYDES AND ACIDS 35 

indicator. The number of c.c. used multiplied 
by the factor 0*0075 gives the fixed acidity in 
terms of tartaric acid in the amount taken. By 
deducting the number of c.c. of decinormal baryta 
solution used for the estimation of fixed acidity from 
the number used for the total acidity (i.e., with the 
25 c.c. before evaporation), a difference is obtained 
which, when multiplied by the factor 0*006, gives the 
volatile acidity as acetic acid. The determination of 
the fixed acidity is important, as some indication is 
thereby given as to whether or not the spirit has been 
stored in sherry-casks. 



32 



CHAPTER V 



THE ESTIMATION OF THE HIGHER 
ALCOHOLS 

THE term " higher alcohols " is convenient, and in- 
cludes, broadly, propyl alcohol, normal butyl alcohol, 
iso-butyl alcohol, and amyl alcohol. The so-called 
fusel oil contains these bodies in varying proportion, 
according to the nature of the spirit and the 
materials from which the spirit has been derived. 
MM. Clandon and Morin and Ordonneau give the 
following as the composition of fusel oil from brandy : 

PERCENTAGE COMPOSITION OF FUSEL OIL ACCORDING TO 
CLANDON AND MORIN AND ORDONNEAU. 





Clandon and 
Morin. 


Ordonneau. 


Propyl alcohol 
Normal butyl alcohol 
Iso-butyl alcohol 
Amyl alcohol 


11-9 
49 '3 
4'5 
34'4 


117 
63-8 

O'O 

24'5 



In the fusel oil of raw potato and grain spirit 
the proportion of amylic alcohol in general pre- 
dominates, that of iso-butyl alcohol next, with 

36 



THE ESTIMATION OF THE HIGHER ALCOHOLS 37 

relatively small amounts of iso-propyl alcohol and 
traces of normal propyl and normal butyl alcohols. 
It is doubtful whether spirits are physiologically bad 
because of the proportion of " higher alcohols " or 
fusel oil which they contain. The amount seldom 
exceeds 0*3 per cent. Age would not appear to 
dimmish this amount, though it is probable that 
the various higher alcohols undergo some not well- 
defined change. Very old brandy often contains 
three times as much higher alcohols than young 
brandy, and ye,t experience is in favour of the former 
being more wholesome than the latter. Several 
processes have been suggested for the determination 
of higher alcohols, some purely of an empirical kind, 
and others of an exact nature. To the former belong 
the colorimetric processes, which, however, give very 
useful and instructive indications, as we shall pre- 
sently see ; to the latter belong those processes based 
on the estimation of the corresponding acids pro- 
duced by oxidizing the alcohols. It is probably due 
to Dupre that we now possess a satisfactory method 
for the determination of higher alcohols, and the 
first intelligible results as to the amount of " fusel 
oil " in brandy were published by him in 1879. 

Dupre's process was improved by Marquardt, who 
extracted the diluted spirit with chloroform, sub- 



38 THE ANALYSIS OF POTABLE SPIRITS 

sequently separating the chloroform and oxidizing 
the higher alcohols contained in it with chromic acid 
mixture. On distillation the acids were obtained, 
and boiled with an excess of carbonate of barium, 
the liquid filtered, evaporated, and the valerate of 
barium weighed. 

A. H. Allen found several inconveniences and 
sources of error in working this process (Journal of 
the Society of Chemical Industry, April 30, 1891), and 
was led to try carbon tetrachloride in the place of 
chloroform for extracting the higher alcohols from 
the diluted spirit. As he points out, this solvent 
has the advantage of boiling at a higher temperature, 
and hence obviating the necessity of oxidizing under 
pressure, and is further obtainable in a state of such 
purity as to yield (contrary to chloroform) no mineral 
acid whatever on oxidation. The distillate, there- 
fore, can be titrated with standard baryta water, 
and the amyl alcohol calculated from the valeric 
acid found, or the solution of valerate of barium can 
be evaporated and the salt weighed. Allen suggests 
certain precautions, as, for example, the addition of 
salt to the spirit, so as to secure complete extraction 
of the amyl alcohol. The following results referring 
to the amount of amylic alcohol in various whiskies 
were obtained by various authors : 



THE ESTIMATION OF THE HIGHER ALCOHOLS 



39 



10 00 to 



i *O O tx PC M M 

; VO M IO fH N <S 



OOOOOOOOOO to to >o 
Ojoppjoppppyoc^Njx 



'^ 



to tx O O 

Tj- to t-t rrj T!- O 






jo p p yoop p ONOO tx fOMD ON O y> p ^t O O O O O O O O ON ON *J 
*O "^i" f*^ tooo ON ff ^ ON ^ oo ^~ O ^O ONVO ^o ^oo ^o ^~ ^O oo CN ONOO oo vo 
^O oo *O ^ N ^* to tx ^ "^ iovo to to 10 Tj~oo 



O ^-.^ too O toTj-O txTj-i-. ONOO txOtoQNOOOOOOOOtxtx 

M to g x O\ M ONVO oo tx tx -< vo O txoo oooooovQN TJ-O\NOO ^J-^O tx N M 

*-^^'-^ ' Kf*PPr*PPr*PPpppppprfprpr*pppp 



i^^ 



ooooooooooooooooooooooooooo 



10 




C/D 



% 0.15 
Pn C rt -*- 1 




^3^x3 2. is :. 

-S-i2.S2888g : 
c^c^c^t> 



a 



co ^ en 



PQ 



40 THE ANALYSIS OF POTABLE SPIRITS 

Allen remarks that it cannot be too clearly under- 
stood that these results represent the apparent fusel 
oil without the separation of the ethers and other 
interfering bodies, and not the actual true amylic 
alcohol, which would be materially lower. He 
believes the results are gravely in excess of the truth, 
owing to the spirit not having been previously treated 
with alkali. P. Schidrowitz employs the following 
modification of this method, which in the author's 
hands has given consistent results : To 100 c.c., or 
preferably 200 c.c., of the spirit 0*15 to 0*2 gramme of 
caustic potash is added, and the whole boiled for an 
hour under the reflux condenser. The alkali saponi- 
fies both the ethers and the furfural. The contents 
of the flask are then transferred to the distilling 
flask, and the distillation carried on as described on 
p. 25. The distillate is diluted with concentrated 
brine until it has a specific gravity of about no. 
It is then extracted with 40, 30, 20, and 10 c.c. re- 
|V j, f spectively of carbon tetrachloride. The combined 
extracts are shaken with 50 c.c. of a saturated solu- 
tion of potassium sulphate, the carbon tetrachloride 
withdrawn, passed through a dry filter, and then 
oxidized with a solution consisting of 30 c.c. water, 
i '5 c.c. H 2 S0 4 (specific gravity, 1*84), and 5 grammes 
of bichromate of potassium, for at least eight hours 



THE ESTIMATION OF THE HIGHER ALCOHOLS 41 

in the water-bath. The whole is then distilled in 
the manner already indicated, first over the naked 

! %4j* It* 

flame of the Bunsen burner, and then with steam^/jkv^ 
until in all about 300 c.c. have passed over. The 
author prefers heating the mixture under a reflux 
condenser for eight hours, and in the distillation it 
is better to place a fresh receiver after the carbon 
tetrachloride has come over. Usually the distilled 
tetrachloride of carbon shows little or no acidity. 
The distillate is finally titrated with a decinormal 
solution of baryta water, first using methyl orange 
as indicator, and secondly phenolphthalein. The 
first result gives any free mineral acid (HC1) and the 
second the valeric acid, which can be calculated into 
amyl alcohol by multiplying the number of c.c. of 
decinormal solution of baryta used, and, after the 
addition of phenolphthalein, by the factor o'Oo88. 
The result may be expressed in terms of iso-butylic 
alcohol by employing the factor 0*0074. 

The colorimetric process for the estimation of 
higher alcohols employed by French chemists, and 
notably those of the Paris Municipal Laboratory, is 
based upon the action of strong sulphuric acid (which 
produces a more or less yellowish-brown colour, 
similar to that obtained by adding Nessler reagent 
to very weak solution of ammonia) upon the higher 



42 THE ANALYSIS OF POTABLE SPIRITS 

alcohols present in alcohol of 50 (in some cases a 
higher) per cent, strength. The method is admit- 
tedly empirical, and cannot be considered as exempt 
from criticism. It is certainly convenient, and gives 
fairly constant and comparative results. Sulphuric 
acid, however, has no action on normal alcohols, 
and the various iso-alcohols show great differences 
of sensitiveness to the action of this reagent. This 
is illustrated in the following table drawn up by 
M. Mohler, showing the relative degree of coloration 
given by the higher alcohols (i in 1,000), in a solution 
of alcohol of 50 per cent. : 

Per Cent. 

Caprylic alcohol - n 

Iso-butylic alcohol - - 10 

(Enanthylic alcohol - "7 

Amylic alcohol - 3 

According to MM. Girard and Cuniasse (" Manuel 
Pratique "), in normal fermentation the higher alco- 
hol which is chiefly present is amylic alcohol, but as 
iso-butylic alcohol is that which is most sensitive to 
the sulphuric acid reaction, they choose this alcohol 
as a standard of comparison for the whole group. 
As furfural and the aldehydes have an action on 
sulphuric acid, these are first eliminated in the fol- 
lowing way : 50 c.c. of the spirit to be tested are 
adjusted to 50 per cent, strength, and placed in a 
flask of a capacity of about 250 c.c. To this quan- 



THE ESTIMATION OF THE HIGHER ALCOHOLS 43 

tity are added 3 grammes of chlorohydrate of meta- 
phenylenediamine or of phosphate of aniline (i c.c. 
of phosphoric acid having a density of 1*453, and 
i c.c. of pure aniline oil). Hewitt's reagent (see 
p. 22) answers this purpose well and most con- 
veniently. A few small pieces of pumice-stone are 
put in to control the ebullition, the flask is then 
attached to a reflux condenser, and its contents 
allowed to boil for one hour, so as to fix the alde- 
hydes. The contents of the flask are then distilled. 
The distillation is allowed to proceed rapidly over 
a naked flame, but not so far as to char the residue. 
The distillate is then brought up to its original 
volume, and the colorimetric tests applied to it. 
For developing the colour, 10 c.c. of the distillate are 
placed in a scrupulously clean flask, and 10 c.c. of 
monohydrated pure and colourless sulphuric acid 
are added by means of a pipette. The acid is allowed 
to trickle slowly from the pipette down the neck of 
the flask, so that it mixes with the alcohol without 
developing any marked degree of heat. The flask 
is next held in a pair of wooden forceps, shaken 
briskly, and heated over the flame of a Bunsen 
burner, care being taken to keep the mixture in 
movement during the heating. As soon as it begins 
to boil, which will be in about fifteen seconds, it is 



44 THE ANALYSIS OF POTABLE SPIRITS 

set aside to cool. Instead of heating the mixture 
directly over a flame, it may be heated in a bath of 
chloride of calcium solution, kept at a temperature 
of 120 C. As soon as the flask is cold, the con- 
tents undergo no further change of colour. The 
same process is carried out with equal volumes of 
acid and the standard solution of iso-butylic alcohol 
(2 per 1,000). The two solutions are matched in the 
colorimeter, and for this purpose the simple appa- 
ratus devised by Hehner is convenient. It consists 
of two graduated cylinders provided with a draw-off 
cock near the foot, so that the depths of the two 
solutions may be brought to show the same intensity 
of colour. MM. Girard and Cuniasse admit that the 
test of coloration developed by sulphuric acid is not 
very delicate, and, in fact, anything below 0*125 
gramme of iso-butylic alcohol per litre scarcely 
shows anything, and consequently an estimation by 
this method in such a case is hardly possible. 
M. Saglier, as a means of increasing the action on the 
higher alcohols, adds a known quantity of furfural 
to both the standard solution and the spirit to be 
tested. Thus, in the case of a spirit evidencing a 
very low amount of higher alcohols there is added 
before heating, both to the specimen to be tested 
and the standard which has been selected for com- 



THE ESTIMATION OF THE HIGHER ALCOHOLS 45 

parison, 20 drops or more of a solution of furfural 
(i in 1,000 in alcohol of 50 per cent, strength). The 
testing is carried out as described above, without 
any attention being paid to the addition of the fur- 
fural. The colour then produced is intensified, and 
has a pinkish tint similar to cobalt nitrate. The 
author finds, however, that its intensity increases 
with exposure to air. 

In order to make the above method more 
sensitive and more exact, M. X. Rocques pro- 
ceeds as follows : 100 c.c. of the spirit to be 
tested, previously distilled and brought to 50 per 
cent, alcoholic strength, are placed in a small flask 
of 250 c.c. capacity. Two grammes of chloro- 
hydrate of metaphenylenediamine (to remove alde- 
hydes and furfurol) and a few pieces of pumice are 
added. The liquid is then boiled gently under a 
reflux condenser for one hour. It is then cooled and 
distilled. The distillation, as before, is done quickly, 
and occupies not more than forty minutes. The 
distillate is received in a flask graduated exactly to 
75 c.c., which will contain the whole of the alcohol, 
and consequently be of a uniform alcoholic strength 
of 667 per cent, after thorough shaking. With a 
pipette 10 c.c. of the distillate are measured into the 
flask of about 100 c.c. measurement, which must be 



46 THE ANALYSIS OF POTABLE SPIRITS 

clean and dry. Then 10 c.c. of monohydrated, 
colourless sulphuric acid are allowed to trickle down 
the neck of the flask. The mixture of acid and 
alcohol is shaken, and then kept at a temperature of 
120 C. for one hour. If many flasks are being tested 
together, it is best to warm them in a calcium 
chloride bath. The calcium chloride solution is kept 
at a constant level by the addition of water, which 
trickles from an inverted flask. The bath is covered 
with a lid furnished with holes of a diameter slightly 
bigger than the test flasks. They are kept upright 
by means of a metal collar which fits round the 
necks and into the holes in the lid. For each series 
of tests it is necessary to provide a comparative 
standard solution which will serve as a colorimetric 
control. The best solution for the purpose is one 
of iso-butylic alcohol containing 0*667 gramme of 
pure iso-butylic alcohol to the litre of alcohol of 
exactly 667 per cent, strength. This solution has 
the same composition as the product of the distilla- 
tion of a solution of 0*500 gramme of iso-butylic 
alcohol in a litre of pure alcohol at 50 per cent, 
strength would have when the distillation has been 
carried as far as experience shows to be the best 
namely, until the original quantity in the retort is 
reduced to one-quarter. In this way the comparison 
between the alcoholic liquid to be tested and the 



THE ESTIMATION OF THE HIGHER ALCOHOLS 47 

standard control solution is made under analogous 
conditions. When the alcoholic liquid to be tested 
and the standard control solution have each been 
submitted to the action of the sulphuric acid for the 
space of one hour at a temperature of 120 C., the 
flasks are withdrawn from the calcium chloride bath 
and are set aside to cool. To compare the depth of 
coloration, M. Rocques employs a Duboscq's colori- 
meter, but Hehner's colorimeter used in Nessler's 
test for ammonia in potable waters is applicable, 
while Mills' colorimeter is excellent for the purpose. 
The intensity of the colour obtained not being 
absolutely proportional to the strength of the liquid 
as regards higher alcohols, the following table giving 
both the apparent strength obtained by the above 
formula with the real strength is adopted by M. 
Rocques for making corrections : 

TABLE III. 

Apparent Strength. Real Strength. 

1*125 1*000 

1*009 0*900 

0'886 . . '. . . . . . 0*800 

0760 . . . . . . . . 0700 

0-640 . . . . . . . . 0*600 

0*500 . . . . . . . . 0*500 

0*379 '4 

0-255 0*300 

0*150 . . . . . . . . 0*200 

O'OOO . . . . . . . . O'lOO 

0*019 0*050 



48 THE ANALYSIS OF POTABLE SPIRITS 

According to the experience of MM. Girard and 
Cuniasse, the coloration obtained by using an alcohol 
containing less than 0*500 gramme of iso-butylic 
alcohol per litre is inappreciable, but this limit of 
i in 20,000 is sufficient for all practical purposes. 
Although it is true that the strengths of 0*5 gramme 
of iso-butylic alcohol in a litre of 50 per cent, alcohol, 
and 0*667 gramme of iso-butylic in a litre of 667 per 
cent, alcohol are the same as regards the percentage 
of higher alcohol per volume of spirit, yet the latter, 
as obtained by distillation, shows nearly double the 
sensitiveness to the test than the former. Thus, the 
solution of 0*667 gramme of iso-butylic alcohol in a 
litre of alcohol at 66*7 per cent, strength obtained by a 
three-fourths distillation gives a colorimetric inten- 
sity of 100, while a solution of iso-butylic alcohol of 
0*500 in a litre of alcohol of 50 per cent, strength only 
gives a colorimetric intensity of 55, or very nearly 
half. For this reason the above procedure is adopted, 
The higher alcohols in ordinary potable spirits are 
completely expelled on distillation, even when they 
exist in a high proportion. As a matter of fact, if a 
solution of iso-butylic alcohol of a strength of 
4 grammes to the litre of alcohol at 50 per cent, be 
taken and distilled, as above directed, the whole of the 
butylic alcohol will be found in the distillate. This is 



THE ESTIMATION OF THE HIGHER ALCOHOLS 49 

in accordance with the author's experience. Accord- 
ing to MM. Girard and Cuniasse, this proportion of 
800 grammes of higher alcohol per hectolitre of 
brandy calculated at 100 per cent, strength may be 
taken to be the maximum which they have ever noted 
in the many analyses of spirituous liquors which they 
have made, and therefore the process described above 
is applicable to any case which may occur in practice. 
Of course, in the case of a very impure alcohol being 
met with, it is necessary to dilute it with a known 
quantity of pure alcohol at 50 per cent, strength 
before proceeding to distil it. In order to obtain 
accurate results, the strength should be strictly com- 
parative, for the coloration produced by the sulphuric 
acid varies, not only with the amount of impurities 
present, but also with the amount of alcohol present. 
The modified method just described gives results 
sufficiently exact to be utilized in the practical 
analysis of potable spirits. M. Rocques remarks 
that in the method of estimation just described the 
colorimetric standard made use of is a solution of pure 
iso-butylic alcohol in ethylic alcohol, and that the 
results are expressed in terms of iso-butylic alcohol. 
This is the standard adopted by the Paris Municipal 
Laboratory. It would appear to be more logical to 
express the results in terms of amylic alcohol, for this 

4 



50 THE ANALYSIS OF POTABLE SPIRITS 

higher alcohol predominates over iso-butylic alcohol 
in brandy and whisky. Amylic alcohol predominates 
also in the fusel oil extracted from the first distilla- 
tion products of potatoes and grain. The results 
which iso-butylic alcohol and amylic alcohol give 
under the same conditions have been investigated 
by MM. Girard and Cuniasse, who carried out their 
experiments upon fermentation alcohols that is to 
say, iso-alcohols obtained by fermentation and 
scrupulously purified by fractional distillation. 
Solutions of these alcohols in pure ethylic alcohol at 
667 per cent, strength were made, and the colora- 
tions given when pure sulphuric acid was added and 
the mixture heated were compared. 

With equal quantities of higher alcohols, the 
coloration produced by amylic alcohol is less than 
that produced by iso-butylic alcohol. The relation 
between the two colorimetric intensities is 6 to 10 
in the experiments made. If iso-butylic alcohol be 
taken as a type, it gives for the higher alcohols 
figures below the truth. Thus, a brandy at 50 per 
cent, alcoholic strength, containing a mean of 
I gramme of higher alcohols per litre, as expressed 
in iso-butylic alcohol, contains really more than 
that amount of amylic alcohol. MM. Girard and 
Cuniasse, however, think that it would not be de- 



THE ESTIMATION OF THE HIGHER ALCOHOLS 51 

sirable to change the standard of comparison hitherto 
used, for it would create a lamentable confusion in 
the analytical results of different observers. On the 
other hand, it is easy to procure a good typical 
standard solution of iso-butylic alcohol, and the 
coloration by this higher alcohol is well denned. 

The action of strong sulphuric acid is employed by 
M. Savalle to estimate the impuretes of spirit in 
the following way : 10 c.c. of alcoholic distillate at 
50 per cent, strength are placed in a small flat- 
bottomed flask with a capacity of 125 c.c. ; 10 c.c. 
of Savalle's reagent pure monohydrated sulphuric 
acid are added. The acid is allowed to trickle down 
the neck of the flask from a pipette, so that it will 
mix with the alcohol without much heating. The 
mixture is then shaken, and heated over the flame 
of a Bunsen burner. As soon as ebullition com- 
mences the flask is removed, covered with a watch- 
glass, and allowed to cool in a place protected from 
dust. When cool, the contents of the flask, which 
are now more or less coloured, are placed in a glass 
tank having parallel faces each 2*5 centimetres 
square. The degree of coloration given by the 
alcohol in this flask is compared with that given by a 
series of glass slips which can be superposed the one 
on the other, forming a chromatic scale numbered 

42 



THE ANALYSIS OF POTABLE SPIRITS 



from o to 15. According to M. Savalle, " o " on his 
scale represents i in 10,000 of impurity. Lovibond's 
tintometer would appear to be admirably adapted 
for making this comparison, but so far the author 
has not tried it. The researches of M. Mohler have 
shown that the sulphuric acid does not act with the 
same intensity upon the various by-products which 
are found in different proportions in alcohol. He 
submitted a mixture of pure alcohol containing a 
proportion of I in 1,000 of the various impuretes 
noted in the following table, and obtained the results 
therein given in degrees of the Savalle scale. 

TABLE IV. 



j Alcohols. 


Aldehydes. 


Ethers. 


Caprylic . . 7 


Iso-butylic . . 9 


Amyl acetate . . .3* 


Iso-butylic . . 6 


Paraldehyde 8 


Ethyl acetate . . .0 


OEnanthylic 4 


Propionic . . 9 


butyrate . o 


Amylic . . 2 


QEnanthylic 5 


Iso-butyrate . o 


Propylic . . o 


Valerianic . . 5 


valerianate . o 


Iso-propylic o 
Butylic . . o 


Ethylic . . 3 '5 
Methylic . . 2*5 


caproate . o 
oenanthylate . o 


Glycerine . . o 


Acetic . . i *5 


sebate . . .0 


Methylic . . o 


Butyric . . o 


succinate . o 




Furfural black 


benzoate . o 






salicylate . o 






formate . . .0 






Methyl salicylate . o 



The following table shows that the coloration 
obtained is not always proportional to the strength 
of the solution of the impurete. 



THE ESTIMATION OF THE HIGHER ALCOHOLS 53 



TABLE V. 



Nature of Solutions. 


Degrees Savalle for Solutions at 


1 00 


Tnnnr 


4 00 


Iso-butylic aldehyde 
Propionic aldehyde 
CEnanthylic alcohol 
Iso-butylic alcohol 
Amyl acetate 


9 
7 

6 
3 


3 
2 '5 
traces 
2-5 
traces 


0-25 
0-25 



0-25 





M. Rocques, using solutions of amylic alcohol in 
ethylic alcohol at 97 per cent, strength, obtained the 
following results, which, as will be seen, are nearly 
proportional : 

A solution of i in 1,000 gave 7 

2 ,, ,, 13 

J> 3 5> ^^ 

The degree of concentration of the alcohol influences 
the results very markedly, as will be seen in Table VI. > 
owing to the greater or less degree of hydration of 
the sulphuric acid. 



TABLE VI. 



Concentration of A Icohol. 


Aldehyde, 
T^nnr- 


Amylic Alcohol, 

TESTS' 


Alcohol at 10 per cent, strength 
50 
70 


0-5 Savalle. 

& : 


0*5 Savalle. 
3'5 
9'0 



54 THE ANALYSIS OF POTABLE SPIRITS 

It is obvious from this table that to get comparable 
results the alcohol used must be of nearly the same 
degree of concentration. The method of heating 
also affects the results, for sometimes the more vola- 
tile impurities evaporate before the heat is suffi- 
ciently great to allow of the carbonizing action of the 
sulphuric acid. MM. Girard and Cuniasse remark 
that this test is very quickly performed, and is 
capable of giving an excellent indication of the 
amount of by-products present in commercial alco- 
hols, while it may be stated that the results given by 
it accord very correctly with the figures by which the 
higher alcohols present in a spirituous liquor are 
estimated. Savalle's test is, of course, empirical, 
but it is a very useful and expeditious means of 
roughly differentiating between malt spirit, grape 
spirit (brandy), and patent spirit, or mixtures of 
patent spirit. The public analyst will find the fol- 
lowing modification by the author of Savalle's test 
very useful as a preliminary guide as to the quantity 
of higher alcohols in a spirit, and in his hands it has 
enabled him to sort quickly all-malt whiskies, 
genuine brandy, etc., from grain spirit or mixtures 
of grain spirit. 

Ten c.c. of the distillate from a spirit with its 
alcoholic strength adjusted exactly to 50 C. are 



THE ESTIMATION OF THE HIGHER ALCOHOLS 55 

taken. To this are added 10 c.c. of monohydrated 
sulphuric acid, specific gravity 1*794 (prepared by 
adding 100 c.c. of pure 1*84 concentrated sulphuric 
acid to 18 c.c. of water and cooling), in a perfectly 
clean test-tube 6 inches long by i inch wide. The 
tube is shaken, the act of mixing, which must be 
completely done, giving rise to development of heat, 
and a small piece of quill-size glass tubing about 
inch long is dropped into the mixture. It is then 
heated over a naked flame, and a note of the time 
made. As soon as a few bubbles of steam escape 
from the fragments of glass tubing, the tube is with- 
drawn from the flame for twenty seconds, and then 
returned, and so on. This method is pursued until 
exactly five minutes have elapsed, when the tube is 
placed in a stream of cold water. When quite cold, 
the contents are placed in a small graduated cylinder, 
and a little alcohol of 50 per cent, strength added to 
restore the volume of the mixture to 20 c.c. Ten c.c. 
of a standard solution of iso-butylic alcohol in 50 per 
cent, alcohol (2 grammes iso-butylic alcohol per litre) 
are treated in exactly the same way in the same test- 
tube. The solution may then be compared, and the 
amount of higher alcohols as iso-butylic alcohol 
calculated. This process gives remarkably consis- 
tent results as long as both the standard solution of 



56 THE ANALYSIS OF POTABLE SPIRITS 

iso-butylic alcohol and the distilled spirit are heated 
in precisely the same way (time intervals, same test- 
tube) and employed at identical alcoholic strengths. 
The results, too, approximate to the amount of 
higher alcohols found by the carbon tetrachloride 
extraction process. By multiplying the amount of 
iso-butylic alcohol formed by the factor 1*19 the 
result may be expressed in terms of amylic alcohol. 



CHAPTER VI 
THE ESTIMATION OF COMPOUND ETHERS 

IT is generally assumed that the chief ethereal salt 
present in brandy and whisky is ethyl acetate, but 
it is known that other esters occur, amongst which 
may be mentioned the compounds of caprylic, capric, 
butyric and caproic acids. On saponifying large 
quantities of the first runnings of the brandy stills 
of Cognac the author was able to prepare several 
ounces of sodium acetate in which very little evi- 
dence was obtained of the existence of higher homo- 
logues. In carrying out the saponification process 
in the laboratory, with the view of determining the 
compound ethers present in spirits, it should be 
borne in mind that aldehyde and furfural are not 
without effect upon caustic alkali. Allen (Journal 
of the Society of Chemical Industry, June 30, 1891) 
finds that the estimation of the ethers from the 
results of their saponification was liable to be 
affected by the presence of bodies such as aldehyde 

57 



58 THE ANALYSIS OF POTABLE SPIRITS 

and furfural, which neutralized more or less alkali. 
Furfural appears to react with a constant ratio of 
alkali according to the following equation : 



2C 5 H A + KOH = KC 5 H 3 3 + C 5 H 6 2 . 

Aldehyde does not appear to give an unvariable 
factor. Allen has therefore suggested the removal 
of aldehydes by boiling the spirit under a reflux 
condenser for two hours with phosphate of aniline 
i c.c. of aniline and i c.c. of phosphoric acid, 
1*442 specific gravity before distillation. The 
author at first adopted this plan, but subsequently, 
with P. Schidrowitz, found that Hewitt's reagent 
was more satisfactory, limiting the boiling under a 
reflux condenser to an hour. In many cases this 
preliminary treatment may be neglected, for the 
" real " and " apparent " figures in regard to 
ether will not differ seriously in the presence of a 
small proportion of aldehyde and furfural. Never- 
theless, it is advisable to prepare sufficient distillate 
free from aldehydes in the manner indicated, both 
for the estimation of higher alcohols and the ethers. 
One hundred c.c. of this distillate is carefully 
neutralized with decinormal baryta solution using 
phenolphthalein as indicator. To this neutralized 
spirit is then added 25 c.c. of an approximately deci- 



THE ESTIMATION OF COMPOUND ETHERS 59 

normal alcoholic solution of soda, and the mixture 
boiled under the reflux condenser for one hour. 
After cooling, the excess of alkali is measured by 
running in decinormal hydrochloric acid until the 
neutral point is exactly reached. The difference 
between the number of c.c. of T ^ HC1 originally re- 
quired for neutralization by 25 c.c. of the alcoholic 
soda solution before saponification and the number 
of c.c. taken to neutralize 25 c.c. after saponifica- 
tion is complete, is an equivalent of the ethers 
present in the quantity of spirit taken. Each c.c. 
represents 0*0088 gramme ethyl acetate. It is 
desirable to carry out a blank experiment with pure 
spirit, and to make any necessary correction arising 
from impurities in the reagents, or possibly from 
the action of the alkali on the glass of the distilling 
flask. Girard and Cuniasse proceed broadly on the 
lines described, but they do not appear to adopt 
special precautions as to the removal of aldehydes 
and furfural before saponification, while they re- 
commend using exactly 10 c.c. of decinormal 
potash for the saponification, then adding when it 
is complete exactly 10 c.c. of decinormal sulphuric 
acid, and finally titrating for the amount of acid in 
excess of the alkali added. 



CHAPTER VII 
ESTIMATION OF FURFURAL 

THE estimation of furfural in spirits is of the utmost 
importance, and fortunately we possess a means of 
determining this body with tolerable accuracy by a 
colorimetric process. Furfural characterizes all 
pot-still spirits ; it is absent in patent spirit, or, if 
present, it is only in relatively small proportion. 
Its presence is due to the action of the open fire on 
the contents of the still during the distillation of 
the wash, and furfural is, in fact, a product of the 
action of heat on cellulose and proteids. It occurs 
in greatest proportion in pot-still whisky ; its 
amount appears to diminish with age, especially 
in brandy, old matured brandy containing generally 
less than immature brandy. The colorimetric 
process for the estimation of furfural is based upon 
the fact that it yields with acetate of aniline a fine 
reddish pink colour. The colour is due to the pro- 
duction of dye-stuffs of a basic character, in which 

60 



ESTIMATION OF FURFURAL 61 

i molecule of the aldehyde unites with 2 mole- 
cules of the base, C 4 H 3 O.CH(C 6 H 2 Me 2 NH 2 ) 2 . Girard 
and Cuniasse employ this reaction for the estimation 
of furfural in spirits in the following manner : 10 c.c. 
of the alcohol to be tested, previously distilled and 
brought up to 50 per cent, alcoholic strength, are 
placed in a graduated tube. In a similar tube are 
placed 10 c.c. of the standard solution of furfural, 
of a strength of 0*005 gramme per litre. To each 
of these tubes is added 10 drops of pure aniline, as 
colourless as possible, and then to both tubes at 
once the time being exactly noted i c.c. of 
acetic acid, pure and free from furfural, is added to 
each. A quarter of an hour after the acid has been 
added the two tubes are compared by means of the 
Duboscq colorimeter. If H be the height of the 
standard solution of furfural, and " h" the height of 
the spirit under test, then this latter will contain 

TT 

0*005 x }T f furfural per litre of alcohol at 50 per 

cent. The test is satisfactory, but certain precau- 
tions must be observed. The aniline oil should be 
as far as possible colourless, and the acetic acid free 
from furfural. P. Schidrowitz (Journal of the 
Society of Chemical Industry, June 30, 1902) states 
that in the case of matured whiskies which he has 



62 THE ANALYSIS OF POTABLE SPIRITS 

examined a yellow colour was developed with the 
aniline reagent. The writer has not experienced 
any difficulty of this kind, but it is necessary that 
the spirit tested should be at exactly 50 per cent, 
alcoholic strength. Schidrowitz has suggested two 
methods of overcoming this contingency by, in the 
first place, using, instead of the ordinary reagent 
5 c.c. aniline and 20 c.c. of glacial acetic acid an 
acetate solution containing a large excess of aniline. 
There is then no yellow colour, but the sensitive- 
ness of the reaction is reduced by one-half. The 
second expedient is to colour the controls. The 
spirit is first decolourized as far as possible by 
adding basic acetate of lead, and removing the 
excess by means of a saturated solution of sulphate 
of potassium. The reagent is then added to the 
sample under examination, and to the control, the 
latter being immediately afterwards coloured with 
a trace of caramel to match the former. This pro- 
cedure, it is stated, does not affect the reaction. 
This process, moreover, overcomes the objection to 
distillation, during the progress of which furfural 
might possibly be formed. The author prefers, how- 
ever, to distil under quite analogous conditions both 
the control furfural solution and the spirit to be 
tested. For this purpose the distillation is allowed to 



ESTIMATION OF FURFURAL 63 

proceed until nine-tenths of the original volume of 
spirit at 50 per cent, strength are collected. It is 
important that the reaction should be permitted to 
proceed for exactly fifteen minutes, and that in all 
cases the alcoholic strength of the control and the 
spirit to be tested is the same. Generally speaking, 
5 c.c. of the distillate of whisky or brandy made up 
to 20 c.c. with 50 per cent, alcohol is a convenient 
quantity to work with, which may be compared in 
two similar graduated tubes with 5 c.c. of the 
standard solution of furfural treated in the same 
way. 

Hewitt proceeds as follows for the estimation of 
furfural (Journal of the Society of Chemical Industry, 
January 31, 1902) : The spirit, if colourless, can be 
employed directly ; if coloured, then a definite 
volume is distilled nearly to the last drop, fresh 
dilute spirit poured into the flask, and the process 
repeated three or four times. The united distillates 
are then made up to some definite volume. Two 
measuring cylinders graduated in c.c., and of the 
same diameter, are chosen. Into one glass 20 c.c. 
of the liquid are poured, into the other is measured 
a quantity of standard furfural solution containing 
approximately the same amount of furfural solution 
as is contained in the 20 c.c. of spirit. The volume 



64 THE ANALYSIS OF POTABLE SPIRITS 

of the standard furfural solution is then also made 
up to 20 c.c. by the dilute spirit. One c.c. of the con- 
centrated aniline acetate solution is added to each 
vessel, and, after standing ten minutes, the tints, on 
looking through the spirits against a white back- 
ground, and also on holding up to the light, are 
carefully adjusted by addition to one or other 
vessel of the dilute alcohol. In making up the 
aniline acetate solution, Hewitt takes equal volumes 
of freshly-distilled aniline, acetic acid, and water, 
boils for a few minutes, and cools to the ordinary 
temperature any furfural present is thus effectu- 
ally destroyed .The standard solutions employed are 
made from freshly-distilled furfural (B. pt. 161 C.), 
and contain 10, i, and 0*1 grammes of furfural in 
i litre of alcohol (specific gravity, 0*920). 

Hewitt gives the following results in milligrammes 
per litre of a number of spirits examined in the way 
indicated : 

SCOTCH WHISKIES. 

1. New malt whisky (Glenlivet type) . . 5*5 

2. Whisky of same distillery (five years 

in cask) . . . . . . . . 4*4 

3. Foreshots of same distillery. . . . 0*3 

4. Moxed low- wines and feints (same 

distillery) O'8 

5. New malt whisky . . . . . . 1*5 

6. 10-5 



ESTIMATION OF FURFURAL 65 

SCOTCH WHISKIES continued. 

7. New malt whisky . . . . . . 1*6 

8. 2-0 

9- 4' 1 

10. 5'5 

11. 6-3 

12. A mature whole-malt whisky (five 

years) . . . . . . . . 4*1 

13. A mature whole-malt whisky (ten 

years) 3'9 

14. A mature whole-malt whisky (Islay, 

seventeen years) . . . . . . 3*2 

OTHER WHISKIES. 

15. New Irish (1900-1901 ; specific 

gravity, 0*8513) . . ' . . . . 6-4 

16. Matured Irish pot-still whisky (? five 

years) 4-5 

17. Canadian club whisky of 1891 . . 1*5 

18. American rye . . . . . . . . 1*4 

VARIOUS SPIRITS. 

19. Brandy (" Fine Old Cognac") .. 17 

20. Old Jamaica rum . . . . . . 3'2 

21. Dop brandy (Cape Colony) . . . . 0*5 

22. (another 
sample) . . . . . . . . 2*1 

23. Cape Smoke . . . . . . . . o'8 

24. Grape brandy (California, 1896) . . 2*2 

25. ( 1898) .. 2-2 

The results, it may be noted, are much higher 
than the amounts of furfural recorded in whisky by 
MM. Girard and Cuniasse, M. Lusson, P. Schidro- 
witz, and the author, a fact which may be ascribed 

5 



66 THE ANALYSIS OF POTABLE SPIRITS 

to the method Hewitt adopts of sweeping out the 
furfural by repeated distillations. It is possible that 
this method of repeated distillations is open to ob- 
jection, owing to the probable formation of fresh 
furfural in the process. By distilling, as previously 
proposed, both the control and the spirit under 
examination down to one-tenth of the original 
volume, distillates are obtained under exactly simi- 
lar conditions of experiment, and the results are 
strictly comparable. The author submitted to dis- 
tillation 100 c.c. of standard furfural 0*01 gramme 
furfural in i litre of 50 per cent, alcohol until the 
distillate measured 90 c.c. It was then made up to 
100 c.c. ; the alcoholic strength was then approxi- 
mately the same as the original standard. On add- 
ing acetic acid and aniline to 5 c.c. of this made up 
to 20 c.c. with alcohol of 50 per cent, strength, the 
shade of colour developed after fifteen minutes was 
exactly the same as that given with 5 c.c. of the 
undistilled standard solution. 

For comparing the tints, Hehner's or Mills' colori- 
meter may be used, or the simple instrument de- 
scribed in the section on colorimeters. 



CHAPTER VIII 
THE IMPORTANCE OF TASTE 

ALTHOUGH, according to the author's experience, 
the results of analysis as detailed in the previous 
chapters will afford data upon which an opinion as 
to whether or not a spirit is what it is described to 
be can be based, yet the analyst should go one step 
further, and cultivate the confirmatory test of taste. 
He should possess a palate, at any rate, which will 
enable him broadly to recognise genuine brandy, 
whisky, and rum. It is obviously a simple matter 
to make up from a plain spirit an alcoholic fluid of 
the same alcoholic strength as genuine malt whisky, 
brandy, or rum, and to add to it artificially-prepared 
by-products ethers, acids, aldehydes, higher alco- 
hols, furfural in such proportion as to make it 
resemble exactly the chemical composition of a 
genuine spirit. Such a mixture, while it would be 
calculated to satisfy the demands of analysis, would 
be at once rejected by the palate. It is, in fact, not 

67 52 



68 THE ANALYSIS OF POTABLE SPIRITS 

possible to imitate the aesthetic qualities of genuine 
spirits by adding by-products to plain spirit, for the 
simple reason that in potable spirits the aldehydes, 
acids, ethers, and higher alcohols are of a very com- 
plex kind, and it is only for convenience of expres- 
sion that these are represented in the simplest terms. 
It is a good plan for the analyst to make up a spirit in 
this way, which should be kept for purposes of refer- 
ence in the laboratory, and to that end a matured 
patent spirit or grain whisky as a spirituous vehicle 
should be chosen. The grain whisky should be 
analyzed as to its by-product contents, and then its 
composition adjusted as regards ethyl acetate, 
furfural, acetic acid, aldehyde, and amyl alcohol, 
according as an imitation brandy, whisky, or rum 
is required for comparison. The sense of smell 
should similarly be cultivated. The plan of allow- 
ing some of the spirit to evaporate on the hands 
and applying the nose to the residue very often 
betrays the presence of foreign admixture. The 
addition of rum to brandy, with a little practice, 
may easily be detected in this way. Analysis will 
confirm the diagnosis by disclosing a proportion of 
ethers inconsistent with other by-products rum 
contains an excessive amount of ethers. In apply- 
ing the sense of taste and smell to whisky, it is best 



THE IMPORTANCE OF TASTE 69 

to dilute the spirit with a little water, which de- 
velops the aroma and flavour of an all-malt spirit, 
while a grain spirit gives little response. Malt 
whisky after a time curiously develops an opal- 
escence when mixed with water, owing to the 
separation of probably oily ethers derived from the 
fats present in the grain. They consist probably 
of ethylic palmitate and oleate. Blended whisky, 
on the other hand, keeps more or less bright when 
water is added to it, according to the amount of 
genuine malt whisky it contains. Pure patent 
spirit remains quite clear. It is possible that by 
carefully measuring the degree of opalescence de- 
veloped on adding a definite volume of water to 
whiskies an approximate estimation of the propor- 
tions of patent spirit and malt spirit could be made. 
For the same reason i.e., that malt whisky con- 
tains oily ethers the spirit, after heating with 
caustic alkali, froths on distillation, a soap being 
formed. A good plan of applying the sense of 
smell to a given spirit is to mix it with some highly- 
charged soda-water. The gas carries with it the 
volatile flavourings, the peculiar characters of which 
a trained nose will readily detect. The difference will 
be found to be very perceptible between a genuine 
old malt spirit and patent spirit when tested in this 



70 THE ANALYSIS OF POTABLE SPIRITS 

way. The former has a " fat," somewhat oily 
smell ; the latter gives only a delicate odour, sug- 
gesting merely wood or storage in a sherry or plain 
cask, as the case may be. 

Old matured malt spirit or brandy, when mixed 
with highly-charged soda-water, " creams " on the 
surface, and the foam is persistent. Patent spirit 
hardly foams at all ; the bubbles break without 
giving a trace of permanent film on the surface. 
Very old brandy mixed with soda-water gives in 
this way a foam as persistent as that of bottled 
beer, and the smell brought out by the escaping 
gas is very characteristic and full of ethereal 
bouquet. 



CHAPTER IX 

GENERAL CONSIDERATIONS OF THE 
RESULTS OF ANALYSIS 

THE French chemists term the by-products of 
potable spirits the impuretes, and the total 
amount of impuretes expressed in milligrammes per 
100 c.c. of alcohol present in the spirit or grammes 
per hectolitre gives the coefficient d? impuretes. 
This coefficient varies widely, not only for different 
spirits, but also for spirits of the same class, accord- 
ing to whether they are young or old. Broadly 
speaking, however, the relationship of the by-pro- 
ducts to each other is preserved, and this relation- 
ship is distinctive of a particular spirit. The total 
amount of secondary products in genuine brandy 
seldom falls below 300 milligrammes per 100 c.c. of 
absolute alcohol present in the spirit. The star 
brandies on the market invariably give figures ap- 
proximating to that sum extreme values are 254 
and 348. The ratio of ethers to higher alcohols in 

71 



72 THE ANALYSIS OF POTABLE SPIRITS 

brandy of the star type is generally unity. Furfural 
is also a very constant factor in brandy, the ordi- 
nary but genuine kinds giving an average of 2 mil- 
ligrammes per 100 c.c. of absolute alcohol present 
in the spirit. In old brandy the acids, ethers, and 
higher alcohols are decidedly increased, but the fur- 
fural is diminished. The total by-products of old 
brandy may reach the figure of 700 milligrammes 
per 100 c.c. of absolute spirit. 

On the other hand, while average Scotch whisky 
shows a coefficient of impuretes similar to average 
brandy, the higher alcohols are greatly in excess of 
the ethers, 3 to i being a common ratio, while fur- 
fural in whisky always exceeds that found in brandy. 
This might be expected having regard to the nature 
of the materials grapes in the one case and barley 
malt in the other from which originally the spirit 
is derived. Part-malt whisky is readily distin- 
guished from all-malt spirit by a sharp fall in the 
coefficient of impuretes, and in all-grain spirit the 
by-products are sometimes almost entirely absent. 
As a rule, however, grain spirit contains some ethers, 
furfural, and higher alcohols, owing partly to storage 
in a sherry cask, and partly to incomplete rectifica- 
tion ; but in any case the amount of each is rela- 
tively small if existent at all (see Table VII.). 



CONSIDERATIONS OF RESULTS OF ANALYSIS 73 

Furfural is rarely present in a grain spirit, and this 
substance affords some basis for estimating the 
extent of admixture of plain spirit with a malt 
whisky or genuine grape-derived brandy. Similarly, 
the ethers may be taken as the basis of calculation, 
more particularly in regard to brandy. Plain spirit 
contains relatively little or no ethers, and its ad- 
mixture with brandy may therefore be estimated 
(see examples given in Table XL). 

In genuine rum the ethers are generally in con- 
siderable excess of the higher alcohols (see Table X.), 
and there is no doubt that rum is used to augment 
the ethers in fictitious brandy (see Table XL) ; 
but the resulting mixture would not contain the 
relative amounts of by-products present in the 
genuine article, while rum would add enormously 
to the volatile acidity. The presence of rum can 
hardly, however, escape the sense of smell or 
palate. By far the greater proportion of the 
whisky sold to the public consists of four-fifths 
grain spirit, and one-fifth genuine malt spirit, 
or the proportion of malt spirit may be^even 
less, or there may be no malt spirit present 
at all. 

Typical blends will be found in Table XL, and 
an approximate idea of the extent of blending may 



74 THE ANALYSIS OF POTABLE SPIRITS 

be gained by comparing the results with those 
obtained with all-malt whiskies in Table VIII. As, 
however, grain whisky invariably contains a small 
proportion of by-products, any calculation based 
upon these would probably in most cases give an 
exaggerated estimate of the real proportion of 
genuine malt spirit. 

Genuine brandies show a much less variable com- 
position than malt whisky, according to the re- 
sults of analyses which are directed to the estima- 
tion of the by-products. Brandy, it should be 
remembered, is sought after for medicinal use on 
account of its stimulating qualities, which are 
superior to any other spirit, and this superiority 
would appear to depend directly upon the invariably 
and relatively high proportion (and kind) of ethers 
which it contains, compared with 'other by-products, 
and especially the higher alcohols. The proportion 
of the latter seldom exceeds the proportion of the 
compound ethers, the ratio of ethers to higher 
alcohols in genuine brandy very often being exactly 
as i is to i, while whisky contains three times as 
much higher alcohols. With genuine rum the reverse 
is the case that is to say, the amount of ethers is as 
much as three times the amount of higher alcohols. 
According to this, rum should be the most powerful 



CONSIDERATIONS OF RESULTS OF ANALYSIS 75 

stimulant we possess, and it is certainly known to 
be a most vigorous restorative, though often a 
source of headache. The ethers in rum, however, 
consist chiefly of ethyl butyrate, which is said to 
have a less favourable physiological action than 
ethyl acetate. 

The fact that brandy is often trusted by the 
public for medicinal purposes, and that, as a rule, 
the genuine article is seldom supplied on draught, 
an altogether inferior and cheaper spirit plain 
spirit, or a mixture of grain spirit flavoured with 
artificial flavours, to which has been added a little 
rum or a little genuine brandy being substituted, 
is ah urgent reason why more control should be 
exercised over the sale of this spirit. 

In a successful prosecution taken under the pro- 
visions of the Sale of Food and Drugs Act at Glas- 
gow, the analysts, Dr. Clark and Dr. J. T. Wilson 
(see Table XI.), based their calculations upon the 
small proportions of ethers found in the specimens 
of brandy submitted to analysis. They regarded 
genuine grape spirit to contain not less than 45 
parts of ether in 100,000 parts of proof spirit (78 
milligrammes per 100 c.c. of absolute alcohol pre- 
sent in the spirit), whereas the sample analyzed 
contained only 13*25 parts per 100,000 parts of 



76 THE ANALYSIS OF POTABLE SPIRITS 

proof spirit (23 milligrammes per 100 c.c. of abso- 
lute alcohol). The taste and smell of the spirit 
accorded with the view that featureless spirit had 
been added. The charge was proven in the Sheriff's 
view, and his judgment was not altered when the 
case was brought forward for appeal in the High 
Court at Edinburgh on December 18, 1903. Several 
successful convictions obtained elsewhere have been 
based on similar reasoning. 

In regard to brandy, M. Lusson observes that the 
oxidation products increase in direct ratio to the 
age of the spirit, but the proportion of the ethers 
and the higher alcohols varies but slightly. It 
follows that the relation of the oxidation products 
to the ethers and the higher alcohols should increase 
in accordance with the age of the brandy. He is 
thus led to estimate the age, and therefore the 
quality, of brandy by what he terms the coefficient 
of oxidation that is to say, by the proportion of 
acids and aldehydes contained in 100 parts of the 
total by-products. The coefficient of oxidation, 
he states, is generally between 10 and 36, being 10 
for young brandies and 36 for a brandy of forty 
years old. The figure rises in exact proportion to 
the age, but shows a constant increase as the age 
of the spirit increases. According to M. Lusson, 



CONSIDERATIONS OF RESULTS OF ANALYSIS 77 

the sum of the ethers and the higher alcohols is 
always more than 0*300 in Cognac brandy which is 
free from any admixture with industrial alcohol. 
It seems to the author that these observations can 
only apply to brandies of unmixed vintages, and 
not to blends. Ageing in cask influences the acidity 
to a certain extent, and not only do the volatile 
acids increase by slow oxidation, but the spirit 
dissolves out certain acid principles from the wood 
which raise the proportion of fixed acids. M. Lusson 
has made some interesting researches on this matter, 
and from the results gained he assumes that he can 
determine the value of a brandy from the point of 
view of its age. 

In regard to the analytical methods employed by 
MM. Girard and Cuniasse, these chemists state that 
the results are generally quite definite enough to 
place an alcohol or a spirituous liquor in the class 
to which it properly belongs, and it is easy to show 
whether an alcohol under examination is a genuine 
spirit or has been adulterated by admixture with 
trade spirit, and, further, whether an alcoholic liquid 
has been artificially flavoured, and whether a trade 
spirit is more or less rectified. 

The methods of analysis of alcohols and spirituous 
liquids adopted by them, and which find a place 



78 THE ANALYSIS OF POTABLE SPIRITS 

amongst others described in these pages, enables 
them to pronounce definitely, they maintain, as to 
the nature of the samples. These methods have 
been practised by them daily in the Municipal 
Laboratory, Paris, for more than ten years, and 
they have always afforded singularly consistent 
results. Moreover, these results have been verified 
by different observers, and their conclusions have 
proved the same. 

Finally, the author has been enabled by the 
methods of analysis described in the text to deduce 
approximately the amount of admixture in various 
spirits, and especially whisky and brandy, submitted 
to him by well-known distillers, who knew accur- 
ately how they were made up, and in every instance 
the deduction was very fairly correct. Moreover, 
when the analytical indications were adjudged to 
be in favour of a spirit being genuine and unmixed, 
the conclusion invariably proved to be in accord- 
ance with fact. 



CHAPTER X 
COLORIMETERS 

THE colorimeter employed by MM. Girard and 
Cuniasse is one devised by Duboscq, which is not 
readily obtainable in this country. The two solu- 
tions are viewed through a disc, one half of which 
receives the light from the standard, and the other 
half from the spirit under examination. The adjust- 
ment is very accurate. 

A useful colorimeter for the analysis of spirits is 
that devised by Mills. It consists of a glass jar 
closed at the top by a cap. The cap is perforated 
in the middle, and carries a short tube in which 
slides a bent glass rod, and supporting exactly at 
right angles to the axis of the jar a flat circular 
opaque white disc. In the bottom of the jar are 
laid a red and green glass disc, one above the other, 
which form a black background to the white disc. 
The white disc is lowered to the bottom of the jar 
by moving its rod, and the jar is filled exactly to 

79 



8o THE ANALYSIS OF POTABLE SPIRITS 

so many divisions with the standard liquid. The 
white disc is raised until on looking vertically down 
through the liquid at the disc the depth of tint 
appears suitable. A second apparatus is then filled 
to the same number of divisions with the liquid to 
be tested, and its disc is adjusted until the colours 
of the columns of the liquid in the two jars appear 
of equal intensity. The lengths of these columns 
are noted, and the comparison is repeated, using 
preferably a different length of column of the 
standard. 

The relative strength of the solutions varies in- 
versely as the lengths of the columns, so that if i c.c. 
of the standard solution correspond with 10 parts 
by weight of the substance estimated, the weight of 
that substance in I c.c. of the test solution will be 

length of column of standard solution multiplied by % 
length of column of test solution 

The Nesslerizing tubes adopted by Hehner are 
still more convenient and easy of application in the 
case of the colorimetric analysis of spirits, but they 
should be of smaller capacity and calibre than those 
used in the estimation of ammonia in waters, while 
the draw-off cock should be near the bottom of the 
cylinder. The two cylinders are placed side by 
side on white paper, and to one of them is added the 



COLORIMETERS 81 

spirit to which the reagent has been added. To the 
other tube a suitable and known amount of standard 
is added. The two liquids are looked at vertically, 
and the darker is allowed to slowly run off into a 
clean beaker until their colours are equal. The 
comparison may be repeated after filling up from 
the beaker. The mode of calculation is the same 
as in former cases. 

The author has been accustomed to employ two 
graduated tubes about i inch in diameter. At the 
lower end is fitted a rubber cork, through the per- 
foration of which passes a glass rod, terminating in 
an opal disc nearly fitting the tube. The discs may 
be pushed up or down until the colours are equal. 
The calculation is the same as before. 



82 



THE ANALYSIS OF POTASLE SPIRITS 



1 1 I-lt^l 
r..^ fr ?5.i?i 

tlliiillsll 

aftfras-gMl^-a 

diH CD O O 



1 
ll 

' 



1 
i 



CO 

8 

tf 

Q 



O O 

. M 

ON w 



OOOO 



COT!- i^ OOOO 

tx'O VO ^OMTj-vo 

co fo fo ro M o\ M txVO M oo 



W 

fc 

H 



CO 

-a-a-a-a 



_0 

r r^ r5 o\ 
O fl fl '^- 



OOOO 
tx p\ M to 

ro In Os M 



OOO 
j-op'^}" 
c^ xooo 



O 

r3r3 XJ ~' 
9 fl eo 






S 



- 





O OOO O O O OOOPQ W 



APPENDIX 



83 



&5 



I 



'pl 



o o 



O ^J 1-1 o3 
^, w O o 



I: 



1 .2 o 











4> rt 0) 

,$ 

.5 -P .2 










Tj-Q'O 

>O "1 O\ 



Ooo ON O 



OO O O O O O oo ON O M O 

O O\ OO -<t fxVO VO Tf N OS tX fO 



O\ N M OO "" 
M "10O uiio 
<t(Si-< r<i 



8O O M O O OOO O O O O O O OO O O 

O O o O O ."t fioo co p O tx O p M 10 ON 

OOOf^O O ON <"*"> O oo O 

O vo N H O oo ON*O vO Th ^J 1 



Ethers as 
Ethyl 
Acetate. 



Tf-Q'Ooooo co OOO O O OOO^i M 

ONMD ON ^vp M p\ p O ^ op covp ^vp ON 

Osvb ^"->b\ "~> MOM ON vb 

Nco^ oo N VOONM vo 



M O 



2.Q N Q POO o o oooco 

w Tj- 5 ^ 5 Tf M Ttoo tx O O ON O ;t 



to rj- O 



o o o o o 

y-> y~> O yo M 
M vo N M rj- 



OOO O O OQQN NO O 

Tt- p c4 co vp y-> p O y> N ON ^- oo 



Description of 



88888 $' $ 



vOoo^Ooooo 



u->Th>-i o 
-MMVO N 



O OOOoo VOO O O 

p p\ ft N op txvp ON N 



a 



BA M*; t^ 

<u fl rt ^ 
c -M -d be M 






a s 












gg 

^^ 



T3 



84 THE ANALYSIS OF POTABLE SPIRITS 




APPENDIX 



85 



g 

o 

Q 







if 

c d 










LJ F-* ^^ fc** 1 '- 

> o > o 



3 ? 3 



X 

w 



CUD o 



St: 



8--a 



s a 



a CO 
ON 



B 2 



O oo 



O O ^ 
co ON -^ 



O co 



fl fl 
O O 



86 THE ANALYSIS OF POTABLE SPIRITS 



d a 
'E. TiL 



a aa -r 



, 
-dig -g ;c 

^ g ^ 

& m 
PQ W 



t 

FP 



ovo 



Q 
W 
Q 
Jz; 
W 
H) 

PQ 



c/) 

W 
P^ 

H 



X 



(3 







o o 



M ON 00 VD 

0\ rj- HH <s 
C^ N N <N 



1 1 



8 8 

1 






u-, O\ tx O O O 
ro *O w vo O\ OO 



oo tx t\ O "~> 



0000 VO X 



41 



w ^t -^ O 
MD Tj- tx O 

O O O"- 1 



O O O >-* '^oo O O ON N 
tx fxMD O tx, O\ N VO >O 



M 1-1 O 



fOO O O 00 

ONOO fo rh 1-1 

>-riVO PO CO N 



rO roOO tx. tx, O 



O\ O M O\ 
tx v> ^N 



O^OOO^Qtx 
Nttvo -^-oo cs O M 



W D O 

/5 S ^ 

K o *- 



W 



ffl 



APPENDIX 



87 



>id 
M^ 

^_i^4d ^^j&t^^-M 
^ <u rf .t) -4-> '33 id 5 

VH O T3 13 fVj-J g ^ d 

a 2 ^^3 .tj ^ g ^'^'S 'S fl 

Ui B C/3 , i C/3 ,_( tuO ^ rf CU D tUD 

O pq pq o W 



Observer. 


d <u 

X 

1 ' ' ' Jj 


British Medic 
Journal 


Do. 

Girard and 
Cuniasse 


^ 


^ I-Q o . 


O\ fo ON f~> O O 
oc JH ON O ON r-1 

Q C^ vo vo Tt" Q 
CO tx OO PO OO O 


o 

oo 


9$ 

vo ON 
txOO 




00 


11 

*"* ^ 


00 QO 

VO VO <^ ^ O Tf 

N OO vo 'rt O t-< 

O O oo oo PO 


O 
ON 


VO 00 

VO - 


a 


e|| 


co vo ro to O O 
ro tx O N Tt - ^- 


& 


Q o 

O oo 


S 


iy 


vo vo CN oo O ON 

O tx tx fO 


2 


VO M 

O tx 


ON 


il 

TT X 


HH ON vo _ O O 

vo * vo " vo C^ 


'd 


j-j 0^ 


8 






vo o O 

vp p O 

O O O 



O O 



O 
ON 



rOp VO 

-*vo ON 









in 



BAILLI&RE, TINDALL AND COX, 
HENRIETTA STREET, COVENT GARDEN, 
LONDON, W.C. 



THIS BOOK IS DUE ON THE LAST DATE 
STAMPED BELOW 



AN INITIAL FINE OF 25 CENTS 

WILL BE ASSESSED FOR FAILURE TO RETURN 
THIS BOOK ON THE DATE DUE. THE PENALTY 
WILL INCREASE TO SO CENTS ON THE FOURTH 
DAY AND TO $1.OO ON THE SEVENTH DAY 
OVERDUE. 



AUG 27 1S35 



AUG 27 1935 



MAY 2 9 1863^ 



| , 



JUN3 1963 



OCT30136594 



REC'D 



mi 3-65-3PM 



LOAN DE.h-1 



LD 2 1-1 1 



889780 



THE UNIVERSITY OF CALIFORNIA LIBRARY